Bayesian approach to rapid In-situ gamma spectroscopy for environmental monitoring

Abstract

In-situ gamma spectroscopy method identifies directly gamma-emitting radionuclides, determines the radioactivity concentration of the identified radionuclides, maintained in the soil, and gives the dose rate estimate without soil sampling from the site. This method can be effectively applied in nuclear accidents or radiological emergencies that require rapid on-site evaluation. A NaI(Tl) detector with high detection efficiency and improved portability is used for on-site measurement, but it has limited low-energy resolution depending on candidate radionuclides in the soil. This study presents an advanced radioactivity estimation method using Bayesian inference that evaluates the uncertainty of radioactivity by counting statistics in the spectrum due to the random nature of radioactive decay. This method is faster, and it gives a more accurate estimation of the radioactivity concentration in the soil than previous methods, even in situations where photo peaks overlap each other due to the low-resolution NaI(Tl) detector. In addition, this study reports the implementation of Gaussian process regression based on a spectrum-to-dose conversion operator. The aim is to deal with uncertainty in dose rate estimation using photon energy and the angle of radiation incidence.